New treatments for people with intellectual disabilities (ID) are increasingly condition-specific in nature. Numerous clinical trials of targeted pharmacological agents are now in process for fragile X syndrome (FXS) and Down syndrome (DS). Condition-specific behavioral treatments also are emerging. Evaluation of all such treatments is being hampered by a lack of adequate cognitive and behavioral endpoints. In this project, we propose to evaluate the adequacy of expressive language sampling for deriving language-relevant clinical endpoints. In this procedure, expressive language samples are collected in highly structured and scripted, yet naturalistic, interactions. These samples can then be analyzed to derive clinical endpoints reflecting important dimensions of language skill and atypical language behavior. Although the suitability of expressive language sampling for clinical trials with FXS or DS (or ID more generally) has yet to be determined, the procedures are especially promising because they yield clinically relevant and functional endpoints, have been shown to capture impairments that are common to ID as well as those specific to FXS or DS, and have been shown to yield robust indicators of developmental change within typical and other language-impaired populations. In this project, we propose: (1) to examine the basic psychometric properties of measures derived from expressive language sampling techniques, including establishing their test-retest reliability, internal consistency, validity, and sensitiviy; (2) to evaluate differences in the psychometric properties of expressive language sampling techniques as a function of variations in participant etiology, age, gender, autism symptom severity, and level of ID; (3) to compare the psychometric properties of three different expressive language sampling techniques; and (4) to evaluate the feasibility of implementing the expressive language sampling across multiple sites, as would be required in a typical clinical trial. These aims will be addressed by collecting expressive language samples from children, adolescents, and young adults with FXS or DS. Samples will be collected within three interaction formats: conversation, narration, and the structured interactions comprising the Autism Diagnostic Observation Schedule (ADOS). Measures derived from the samples will include those indexing syntax (an area of especially severe impairment in DS) and perseveration (an area of especially severe impairment for FXS). Test-retest reliability will be assessed at 4 weeks (+/- 1 week) using alternate versions of sampling materials. Internal consistency will be assessed by computing alpha coefficients within and across sampling techniques. Standardized tests and informant report will be used as indicators of validity. A two-year longitudinal follow-up will yield an estimate of sensitivity to change. Participants will be tested at multiple sites, each with considerable experience in the evaluation of individuals with FXS or DS. Feasibility of multiple-site implementation will be evaluated by comparing language samples across sites on key indicators. Transcription, coding, and analysis will be conducted only at the UC Davis MIND Institute site.